Excess Mortality Rate Associated with Hepatitis C Virus Infection: a Community-Based Cohort Study in Rural Egypt

Research Article

Excess mortality rate associated with hepatitis C virus infection: A community-based cohort study in rural Egypt

Aya Mostafa1,y, Yusuke Shimakawa2,y, Ahmed Medhat3, Nabiel N. Mikhail4, Cédric B. Chesnais2,5, Naglaa Arafa1, Iman Bakr1, Mostafa El Hoseiny1, Mai El-Daly6,7, Gamal Esmat8, ⇑ Mohamed Abdel-Hamid6,9, Mostafa K. Mohamed1, Arnaud Fontanet2,10,

1Department of Community, Environmental and Occupational Medicine, Faculty of Medicine, Ain Shams University, Cairo, Egypt; 2Unité d’Épidémiologie des Maladies Émergentes, Institut Pasteur, Paris, France; 3Department of Gastroenterology & Tropical Medicine, Faculty of Medicine, Assiut University, Assiut, Egypt; 4Department of Biostatistics and Cancer Epidemiology, South Egypt Cancer Institute, Assiut University, Assiut, Egypt; 5UMI 233, Institut de Recherche pour le Développement (IRD), Montpellier, France; 6Viral Hepatitis Research Laboratory, National Hepatology and Tropical Medicine Research Institute, Cairo, Egypt; 7National Liver Institute, Menouﬁa University, Menouﬁa, Egypt; 8Endemic Medicine Department, Faculty of Medicine, Cairo University, Cairo, Egypt; 9Department of Microbiology and Immunology, Faculty of Medicine, Minia University, Minia, Egypt; 10Département d’Infection et Épidémiologie, Conservatoire National des Arts et Métiers, Paris, France

Background & Aims: >80% of people chronically infected with Conclusions: Use of a highly potent new antiviral agent to treat hepatitis C virus (HCV) live in resource-limited countries, yet all villagers with positive HCV RNA may reduce all-cause mortal- the excess mortality associated with HCV infection in these set- ity rate by up to 5% and hepatic mortality by up to 40% in rural tings is poorly documented. Egypt. Methods: Individuals were recruited from three villages in rural Ó 2016 European Association for the Study of the Liver. Published Egypt in 1997–2003 and their vital status was determined in by Elsevier B.V. All rights reserved. 2008–2009. Mortality rates across the cohorts were compared according to HCV status: chronic HCV infection (anti-HCV antibody positive and HCV RNA positive), cleared HCV infection Introduction (anti-HCV antibody positive and HCV RNA negative) and never infected (anti-HCV antibody negative). Data related to cause of Hepatitis C virus (HCV) infection is an important public health death was collected from a death registry in one village. issue. Globally, 115 million people have been infected with Results: Among 18,111 survey participants enrolled in 1997– HCV, of whom 80 million are chronically infected [1]. Each year, 2003, 9.1% had chronic HCV infection, 5.5% had cleared HCV an estimated 700,000 people die from HCV, mainly through liver infection, and 85.4% had never been infected. After a mean time cirrhosis and/or hepatocellular carcinoma (HCC) [2]. The majority to follow-up of 8.6 years, vital status was obtained for 16,282 (>80%) of people chronically infected with HCV reside in low- and (89.9%) participants. When compared to those who had never middle-income countries [1], and Egypt has the highest preva- been infected with HCV in the same age groups, mortality rate lence of chronic HCV infection, estimated at 10.0% in adults [3]. ratios (MRR) of males with chronic HCV infection aged <35, With the recent advent of new antiviral therapy which is 35–44, and 45–54 years were 2.35 (95% CI 1.00–5.49), 2.87 expected to cure more than 90% of chronic infection within a (1.46–5.63), and 2.22 (1.29–3.81), respectively. No difference in short duration [4], Egypt has started treating a large number of mortality rate was seen in older males or in females. The infected persons through 23 national treatment centres with a all-cause mortality rate attributable to chronic HCV infection regimen that includes the new antiviral agent sofosbuvir [5].To was 5.7% (95% CI: 1.0–10.1%), while liver-related mortality was assess the population impact of this national treatment pro- 45.5% (11.3–66.4%). gramme, it is critical to understand the natural history of chronic HCV infection and the excess mortality associated with chronic HCV infection in the local context before treatment programmes Keywords: Hepatitis C; Mortality; Cohort studies; Egypt; Africa. are implemented. Received 6 October 2015; received in revised form 12 February 2016; accepted 16 Several cohort studies have examined the impact of HCV February 2016; available online 26 February 2016 infection on all-cause and cause-specific mortality rates in ⇑ Corresponding author. Address: Unité d’Épidémiologie des Maladies Émergen- general population [6–16]. However, these studies have been tes, Institut Pasteur, 25 rue du Docteur Roux, Paris, France. Tel.: +33 1 4061 3763; fax: +33 1 4568 8876. restricted to high income countries. To date there has been no E-mail address: [email protected] (A. Fontanet). similar study conducted in middle- and low-income countries. y These authors have contributed equally as joint first authors. Moreover, the estimates from studies conducted in high income Abbreviations: HCV, Hepatitis C Virus; HCC, Hepatocellular carcinoma; HIV, countries have limited generalisability to other settings. For Human Immunodeficiency Virus; HBV, Hepatitis B Virus; RNA, Ribonucleic acid; example, among people with chronic HCV infection, those who PAF, Population attributable fraction; MRR, Mortality rate ratios; PY, Person- years; 95% CI, 95% confidence interval; IQR, Interquartile range. inject drugs – a major route of HCV transmission in Europe and

Journal of Hepatology 2016 vol. 64 j 1240–1246 JOURNAL OF HEPATOLOGY

North America – are thought to have additional causes of of death was classified into one of eight categories: liver-related (including increased mortality, such as drug overdose, excessive alcohol HCC), neoplasms (excluding HCC), stroke, heart disease, pulmonary disease, kidney disease, other and unknown causes [11]. intake, and co-infection with HIV or hepatitis B virus (HBV) [7–9,13,14]. In contrast, in resource-limited countries, iatrogenic Statistical analyses procedures are the most frequent mode of HCV transmission [17]. In Egypt, the beginning of the HCV epidemic has been asso- HCV status at enrolment was treated as a categorical variable (chronically ciated with the mass parenteral anti-schistosomal treatment infected, cleared HCV infection and never infected). Baseline characteristics of campaigns in the 1960s–70s [18], and has since spread through participants were compared according to each HCV status using Chi-square test the reuse of medical devices [19]. As a result, in a generalized for categorical variables and Kruskal-Wallis one-way analysis of variance for con- epidemic, such as the one of Egypt, it is expected that the risk tinuous variables. The person-years of follow-up were calculated from the date the participants were enrolled in the study to the date of death, migration, or last of mortality among HCV-infected in Egypt will be different to that follow-up, whichever came first. The association between HCV status and all- found in high income countries. cause mortality was examined by estimating mortality rate ratios (MRR) using We therefore conducted a community-based cohort study in Poisson regression, and adjusted for potential confounding variables: sex, current rural Egypt to estimate the all-cause mortality rates according age, study village, education (ever or never attended school: a proxy for socio- to HCV status: chronic HCV infection (anti-HCV antibody positive economic status), and comorbidity. Comorbidity was defined as the presence of any of the following medical history at enrolment: P2 episodes of blood transfu- and HCV RNA positive), cleared HCV infection (anti-HCV antibody sion, P2 hospital admissions, P2 surgical procedures or undertaking dialysis. As positive and HCV RNA negative) and never infected (anti-HCV HCV-related mortality risk is known to be higher in males compared to females antibody negative). In addition, we estimated the fraction of and in older persons compared to other age groups [8], the interactions between all-cause and liver-related mortality rates attributable to chronic sex and chronic HCV infection, and between age and chronic HCV infection, on mortality were examined using a likelihood ratio test by adding interaction terms HCV infection (population attributable fraction: PAF). in a Poisson regression model adjusted for other confounders (study village, education and comorbidity). The cohort in Zawiat Razin was used to describe the cause of death, and the association between HCV status and liver-related mortality. Finally, the PAF and its 95% confidence interval (95% CI) were estimated for Patients and methods the effect of chronic HCV infection on all-cause mortality rate (across all three villages) and on liver-related mortality rate (in Zawiat Razin). The ‘‘punaf” STATA Baseline sero-surveys command was used after fitting a Poisson regression model that included potential confounders and the effect modifiers [29]. All analyses were performed using Community-based sero-surveys for HCV infection were conducted in three rural STATA 13.0 (Stata Corporation, College Station, TX). villages in Egypt. In Sallam, Upper Egypt, all villagers aged P5 years were invited to participate in the sero-survey between 1997–1999 [20–22]. In two villages in the Nile Delta, Aghour El Soughra [22–24] and Zawiat Razin [19,25,26], randomly Results selected inhabitants aged P5 years were invited to the surveys between 1997– 1999 and 2002–2003, respectively. After written informed consent was provided, Baseline characteristics participants underwent a standardised clinical examination, a blood sample was collected, and a structured interview that assessed socio-demographic status and potential risk factors for HCV transmission was conducted. Participants identified A total of 18,111 inhabitants participated in the baseline sero- as having a chronic HCV infection were referred to a hepatology clinic for further surveys (4,311 in Aghour El Soughra, 7,385 in Sallam and 6,415 clinical management, and those eligible and willing to be treated were offered in Zawiat Razin) in 1997–2003. Survey uptake was 62.8%, 75.4% pegylated interferon (PegIFN) and ribavirin [27]. The study was approved by the Institutional Review Board of the Egyptian Ministry of Health and Population, and 77.2%, the prevalence of positive anti-HCV antibodies was the Assiut University (Assiut, Egypt), and the University of Maryland (Baltimore, 8.7%, 24.3% and 11.8%, and the prevalence of positive HCV RNA USA). in those sero-positive for anti-HCV antibodies was 63.0%, 65.5% and 59.9%, in Sallam [20], Aghour El Soughra [23] and Zawiat Laboratory assays Razin [19], respectively. At the follow-up visit in 2008–2009, vital status was ascer- Serum samples were transferred and tested at the National Hepatology and Trop- tained in 89.9% (16,282) of the initial survey participants. No sig- ical Medicine Research Institute in Cairo. Sera from Sallam and Aghour El Soughra nificant difference was observed in the distribution of HCV status were tested for anti-HCV antibody using a second-generation enzyme immunoas- between persons whose vital status was ascertained and persons say (Abbott HCV EIA 2.0, Abbott Laboratories, Chicago, IL) and positive samples who were lost to follow-up. The baseline characteristics of these were confirmed using a recombinant immunoblot assay (RIBA, Chiron, Emery- ville, CA). Sera from Zawiat Razin were tested using a third-generation anti- 16,282 participants are presented in Table 1. At the time of enrol- HCV assay (INNOTEST HCV Ab IV, Innogenetics, Ghent, Belgium), followed by a ment, 1,647 (9.1%) were found to be chronically infected with confirmation test using AxSYM HCV version 3.0 (Abbott Laboratories). All sera HCV, 996 (5.5%) to have cleared previous HCV infection and positive for anti-HCV antibodies were further tested for HCV RNA using a direct 15,468 (85.4%) had never been infected with HCV (Fig. 1). Com- in-house reverse transcriptase polymerase chain reaction [28]. pared to those never infected with HCV (n = 13,924), people with chronic HCV infection (n = 1,473) were more likely to be older, of Ascertainment of death male gender, to have never attended school, and to have comorbidity. Prevalence of chronic HCV infection also differed signifi- Between 2008 and 2009, a team of fieldworkers visited each household in the cantly among the villages: Aghour El Soughra (16.8%), Zawiat study villages and interviewed household heads to determine vital status of all Razin (8.2%), and Sallam (5.7%) (p <0.001). the survey participants originally recruited between 1997 and 2003. In case of migration of study participants during this period, the date of migration was recorded. In the case of death of a study participant during the follow-up period, Crude all-cause mortality the date of death was ascertained through the death registry held in the primary healthcare centre in each village. If this information was not available through the death registry, the date reported by family members was recorded. In Zawiat During a mean time to follow-up of 8.6 years (139,809 person- Razin, the cause of death was also recorded from the death registry. The cause years (PY)), 591 (3.6%) died, resulting in an overall mortality rate

Journal of Hepatology 2016 vol. 64 j 1240–1246 1241 Research Article

Table 1. Baseline characteristics of 16,282 participants according to HCV infection status.

Total Chronic HCV infection Cleared HCV infection Never infected p value (n = 16,282) (n = 1473) (n = 885) (n = 13,924) Median age (years, IQR) 22 (12-36) 40 (30-50) 38 (25-50) 19 (12-32) <0.001 Sex (n, %) Female 8326 (51.1) 632 (42.9) 462 (52.2) 7232 (51.9) <0.001 Male 7956 (48.9) 841 (57.1) 423 (47.8) 6692 (48.1) School Ever attended 7536 (48.7) 541 (37.6) 312 (36.2) 6683 (50.8) <0.001 Never attended 7929 (51.3) 898 (62.4) 549 (63.8) 6482 (49.2) Village Sallam 7127 (43.8) 408 (27.7) 267 (30.2) 6452 (46.3) <0.001 Aghour El Soughra 3655 (22.4) 615 (41.8) 358 (40.4) 2682 (19.3) Zawiat Razin 5500 (33.8) 450 (30.5) 260 (29.4) 4790 (34.4) Blood transfusion 0-1 16,162 (99.3) 1438 (97.6) 862 (97.4) 13,862 (99.6) <0.001 ≥2 119 (0.7) 35 (2.4) 23 (2.6) 61 (0.4) Hospital admission 0-1 15,361 (94.3) 1339 (90.9) 801 (90.5) 13,221 (95.0) <0.001 ≥2 920 (5.7) 134 (9.1) 84 (9.5) 702 (5.0) Surgical procedure 0-1 15,097 (92.7) 1284 (87.2) 748 (84.5) 13,065 (93.8) <0.001 ≥2 1184 (7.3) 189 (12.8) 137 (15.5) 858 (6.2) Dialysis No 16,263 (99.9) 1468 (99.7) 882 (99.7) 13,913 (99.9) 0.001 Yes 18 (0.1) 5 (0.3) 3 (0.3) 10 (0.1) Comorbidity1 No 14,589 (89.6) 1210 (82.2) 713 (80.6) 12,666 (91.0) <0.001 Yes 1693 (10.4) 263 (17.8) 172 (19.4) 1258 (9.0)

1Comorbidity is deﬁned as the presence of any of the following medical history: P2 episodes of blood transfusion, P2 hospital admissions, P2 surgical procedures or dialysis.

Total no. screened N = 18,111 Aghour El Soughra (1997-1999): n = 4311 Sallam (1997-1999): n = 7385 Zawiat Razin (2002-2003): n = 6415

Chronic HCV at enrolment Cleared HCV at enrolment Never infected at enrolment N = 1647 (9.1%) N = 996 (5.5%) N = 15,468 (85.4%)

Lost to follow-up Lost to follow-up Lost to follow-up N = 174 (10.6%) N = 111 (11.1%) N = 1544 (10.0%)

No. followed in 2008 No. followed in 2008 No. followed in 2009 N = 1473 N = 885 N = 13,924

Migrated: n = 91 (6.2%) Migrated: n = 46 (5.2%) Migrated: n = 1664 (12.0%) Alive: n = 1247 (84.6%) Alive: n = 776 (87.7%) Alive: n = 11,867 (85.2%) Death: n = 135 (9.2%) Death: n = 63 (7.1%) Death: n = 393 (2.8%)

Fig. 1. Flow diagram of study participants.

of 4.2 (95% CI: 3.9–4.6) per 1,000 PY. Across this period, 1,801 fore, only estimates adjusted for confounding factors (among (11.1%) participants migrated (Fig. 1). Participants with chronic which age and sex) should be considered for meaningful interpre- HCV infection had the highest crude mortality rate (10.4, 95% tation of the results. CI: 8.8–12.3), followed by those with cleared HCV infection (8.0, 6.2–10.2) and those who were never infected with HCV Adjusted all-cause mortality (3.3, 3.0–3.6 per 1,000 PY) (Table 2). However, the higher crude mortality rate observed in subjects with chronic HCV infection After adjusting for sex, current age, education, village and comor- may be accounted by their older age and higher male proportion, bidity, the mortality rate of those chronically infected with HCV particularly when compared to the non-infected group. There- was signiﬁcantly higher than that of never infected (MRR: 1.26,

1242 Journal of Hepatology 2016 vol. 64 j 1240–1246 JOURNAL OF HEPATOLOGY

95% CI: 1.02–1.56, p = 0.03). There was no difference between Table 3. Mortality rate ratios for patients with chronic HCV infection those who cleared HCV and those who were never infected compared to never infected, stratified by age group and sex. (MRR: 0.91, 95% CI: 0.69–1.21, p = 0.5). Older age, villages other Age group (yr) Men Women than Sallam, and the presence of comorbidity were also associ- Adjusted MRR1 p value Adjusted MRR1 p value ated with higher all-cause mortality rate (Table 2). <35 2.35 (1.00-5.49) 0.05 1.27 (0.53-3.07) 0.6 35-44 2.87 (1.46-5.63) 0.002 1.55 (0.77-3.13) 0.2 Age and sex modify the effect of chronic HCV infection on mortality 45-54 2.22 (1.29-3.81) 0.004 1.20 (0.67-2.14) 0.5 55-64 1.45 (0.92-2.27) 0.1 0.78 (0.47-1.29) 0.3 There was strong evidence that sex modifies the effect of chronic ≥65 1.32 (0.94-1.85) 0.1 0.72 (0.47-1.10) 0.1 HCV infection on risk of all-cause mortality as compared to those 1 who had never been infected with HCV: the adjusted MRR was Adjusted for education, village and comorbidity. 1.55 (95% CI: 1.19–2.02) in men and 0.86 (0.60–1.23) in women (p = 0.009, likelihood ratio test). The estimates of the association between chronic HCV infection and all-cause death also found a cardiac deaths (18.7%). Among people with chronic HCV infection significant modifying effect of age (p = 0.008). Moreover, inclu- (n = 30), liver-related death was the most frequent cause (36.7%), sion of both chronic HCV infection and sex, and then chronic followed by cardiac death (10.0%). In contrast, in people who HCV and age groups in the model provided significantly better cleared HCV infection (n = 15) and in those who had never been fit than including either of the interaction terms alone. Conse- infected (n = 110), cardiac death was the leading cause of death. quently, all-cause MRRs comparing patients with chronic HCV Compared to people who died from other causes, those who died infection to those who had never been infected, adjusted for edu- of liver-related causes had a younger median age (52 years, cation, village and comorbidity, are presented in ten subgroups interquartile range (IQR): 43–61 vs. 63, IQR: 49–72, p = 0.02) stratified by sex and current age groups (Table 3). Among men, and were more likely to be men (65.0% vs. 50.4%, p = 0.2). the adjusted MRR was 2.35 (95% CI: 1.00–5.49) in those aged All-cause and cause-specific MRRs in Zawiat Razin are pre- <35 years, 2.87 (1.46–5.63) in 35–44 years old, 2.22 (1.29–3.81) sented in Table 5. After adjusting for sex, age, education and in 45–54 years old, 1.45 (0.92–2.27) in 55–64 years old and comorbidity, the strength of association between chronic HCV 1.32 (0.94–1.85) in those aged P65 years. In women, the age- infection and all-cause mortality in Zawiat Razin (MRR: 1.23) specific MRRs were close to unity across the age groups. was similar to that observed across all three villages (MRR: 1.26). However, the increase in risk in Zawiat Razin was not sta- Liver-related mortality tistically significant due to the lack of power with data from one village only: using never infected as a reference, the MRR was Of the 5,500 people from the village of Zawiat Razin whose vital 1.02 (95% CI: 0.59–1.77) in those with cleared HCV infection status was recorded, 155 people died after mean time to follow- and 1.23 (0.81–1.86) in those with chronic HCV infection. The up of 6.1 years. Classified cause of death is presented in Table 4. results were similar in a model that excluded comorbidity as a Overall, 13.6% of deaths were liver-related, ranking second after covariate. In contrast, the incidence of liver-related death was

Table 2. Factors associated with all-cause mortality.

Variables PY No. of deaths Rate per 1000 PY (95% CI) Adjusted MRR1 MRR (95% CI) p value HCV Never infected 118,961 393 3.3 (3.0-3.6) Reference 0.04 Cleared 7882 63 8.0 (6.2-10.2) 0.9 (0.7-1.2) Chronic infection 12,966 135 10.4 (8.8-12.3) 1.3 (1.0-1.6) Sex Female 70,006 283 4.0 (3.6-4.5) Reference 0.1 Male 69,803 308 4.4 (3.9-4.9) 1.2 (1.0-1.4) Current age <15 26,705 12 0.4 (0.3-0.8) Reference <0.0012 (years) 15-24 40,173 33 0.8 (0.6-1.2) 1.8 (0.9-3.6) 25-34 25,954 34 1.3 (0.9-1.8) 2.7 (1.4-5.3) 35-44 19,663 48 2.4 (1.8-3.2) 4.9 (2.6-9.3) 45-54 13,565 74 5.5 (4.3-6.9) 9.7 (5.2-18.0) 55-64 8365 127 15.2 (12.8-18.1) 28.6 (15.7-52.2) ≥65 5386 263 48.8 (43.3-55.1) 91.3 (50.7-164.5) School Ever attended 67,964 133 2.0 (1.7-2.3) Reference 0.2 Never attended 64,789 428 6.6 (6.0-7.3) 1.2 (0.9-1.5) Village Sallam 68,515 226 3.3 (2.9-3.8) Reference 0.001 Aghour El Soughra 37,655 210 5.6 (4.9-6.4) 1.6 (1.3-1.9) Zawiat Razin 33,638 155 4.6 (3.9-5.4) 1.3 (1.0-1.6) Comorbidity No 125,853 471 3.7 (3.4-4.1) Reference 0.003 Yes 13,955 120 8.6 (7.2-10.3) 1.4 (1.1-1.7)

1Adjusted for HCV status, sex, current age, education, village and comorbidity; 2Test for linear trend.

Journal of Hepatology 2016 vol. 64 j 1240–1246 1243 Research Article

Table 4. Cause of death by HCV status in Zawiat Razin (n = 155). liver-related mortality rate but similar overall mortality; and iv) 5.7% of all-cause and 45.5% of liver-related mortality rates were Total Chronic Cleared Never HCV HCV infected attributable to chronic HCV infection. (n = 155) (n = 30) (n = 15) (n = 110) Similar to previous community-based cohort studies in indus- Liver-related 20 (12.9%) 11 (36.7%) 3 (20.0%) 6 (5.4%) trialized countries [12–14,16], we found an increased all-cause Neoplasms excluding 8 (5.2%) 1 (3.3%) 2 (13.3%) 5 (4.6%) mortality rate in people with chronic HCV infection compared HCC to people never infected with HCV. The effect of chronic HCV Stroke 16 (10.3%) 2 (6.7%) 0 14 (12.7%) infection on all-cause mortality differed significantly between Heart disease 30 (19.4%) 3 (10.0%) 4 (26.7%) 23 (20.9%) sex and between age groups, with this effect of increased mortal- Pulmonary disease 7 (4.5%) 1 (3.3%) 3 (20.0%) 3 (2.7%) ity only observed in men up to the age of 54 years. Further anal- Kidney disease 10 (6.4%) 1 (3.3%) 1 (6.7%) 8 (7.2%) ysis in the Zawiat Razin cohort suggests that this excess in all- Other 50 (32.3%) 7 (23.3%) 0 43 (39.1%) cause deaths was mainly due to an increased liver-related mor- Unknown 14 (9.0%) 4 (13.3%) 2 (13.3%) 8 (7.3%) tality in chronically infected individuals. In this cohort, the median age of liver-related deaths (52 years) was lower compared to that of other causes of death (63 years), implying that in rural Egypt, the relative burden of chronic HCV infection is most important in young and middle-aged men, rather than in older Table 5. Adjusted cause-specific mortality rate ratios by HCV status in Zawiat men. One explanation may be that in older men, the increased 1 Razin (n = 5,500). frequency of cardiac or cerebrovascular events may be diluting Causes Adjusted MRRs (95% CI) p value the mortality risk associated with chronic HCV infection. There Never Cleared HCV Chronic HCV was no increase in mortality rate associated with chronic HCV infected infection in women when compared to women who had never (reference) been infected. It is widely accepted that the speed of progression All-cause 1.00 1.02 (0.59-1.77)1.23 (0.81-1.86)0.6 from asymptomatic chronic HCV infection to liver cirrhosis, and Hepatic 1.00 3.83 (0.95-15.47)7.75 (2.81-21.40)<0.001 from liver cirrhosis to death is slower in women than men [30– diseases 32], which may explain our finding. Still, it is expected that with Extrahepatic 1.00 0.77 (0.40-1.48)0.70 (0.40-1.23)0.4 additional years of follow-up, an increase in HCV-related mortal- diseases ity will emerge in women as well. Unknown 1.00 1.82 (0.37-8.91)2.04 (0.58-7.20)0.5 In contrast to participants with chronic HCV infection, those cause who cleared HCV infection had a similar all-cause mortality rate 1Adjusted for sex, current age, education and comorbidity. compared to those never infected in this study. This differs from reports from Western countries where injecting drug use is the major mode of HCV transmission and people who cleared HCV significantly higher among those with cleared HCV infection share the same lifestyle risks (drug overdose, suicide, murder, (adjusted MRR: 3.83 (95% CI: 0.95–15.47)) and those with chronic excessive alcohol consumption, HIV co-infection) as those who HCV infection (adjusted MRR: 7.75 (95% CI: 2.81–21.40)). There became chronically infected with HCV [8,12,14]. In Denmark, was no evidence of interaction between HCV status and sex, all-cause mortality of people who cleared HCV was significantly nor between HCV status and age on hepatic mortality. Non- higher than that of never infected, except in a subgroup of indi- liver-related mortality rates and death rates with unknown cause viduals aged 40–69 years without injecting drug use, alcohol did not differ according to HCV status. abuse and other comorbidities, implying that in the absence of high-risk behaviours there is little excess mortality in people Population attributable fractions (PAF) who cleared HCV infection [14]. However, we found that liver- related mortality was higher in people who cleared HCV infection After fitting a Poisson regression model that included education, compared to never infected. This association is not unique to our village, comorbidity and interactions between HCV infection and study and has been observed in previous community-based sex and between HCV and age, 5.7% (95% CI: 1.0–10.1%) of all- cohort studies [12,14,16]. This may be explained either by a gen- cause mortality rate was attributable to chronic HCV infection. uine progression of chronic liver disease despite clearance of the In Zawiat Razin, 45.5% (95% CI: 11.3–66.4%) of liver-related mor- virus in few patients who had advanced fibrosis and who were tality was attributable to chronic HCV infection. subsequently treated, or by a misclassification bias, in which the knowledge of anti-HCV positivity may influence the diagnosis associated with the cause of death. Discussion The findings of this study may have been limited by the length of follow-up and potential misclassification of exposure. The risk This is the first longitudinal community-based cohort study of advanced liver disease has previously been found to increase assessing the effect of HCV infection on mortality in a resource- 10–20 years after the original infection [33]. If, for example, most limited setting. In rural Egyptian villages, we found that: i) of the villagers in these cohorts were infected after 1995, they chronic HCV infection was associated with all-cause mortality would not have yet reached the period of elevated HCV-related in young and middle-aged men, but not in women or older mortality. Although we do not know the age at which partici- men; ii) chronic HCV infection was a risk factor for hepatic death pants were infected, this is an unlikely explanation as it is irrespective of sex and age group; iii) compared to people never thought that a substantial number of each cohort were infected infected, those who cleared HCV infection had an increased much earlier (at least 12.4% of HCV infection in Zawiat Razin in

1244 Journal of Hepatology 2016 vol. 64 j 1240–1246 JOURNAL OF HEPATOLOGY

2002 was attributable to schistosomiasis intravenous treatment Financial support programme dated in 1960–1980 [19]). Further, misclassification of HCV status may have occurred as HCV transmission is still The cohort in Zawiat Razin was funded by ‘‘Agence Nationale de ongoing in the villages included in the study, with estimates of Recherche sur le SIDA et les Hépatites Virales” (ANRS 1211 & 0.1–0.7% of those never infected with HCV at enrolment acquiring 1279). The cohort in Sallam was funded by U.S. Agency for Inter- HCV infection each year during the study period [22,25]. In addi- national Development (263-G-00-96-00043), National Institute tion, 10% of the patients with chronic HCV infection agreed to of Allergy and Infectious Diseases/National Institute of Childhood receive combined PegIFN and ribavirin as part of their participa- Diseases (5U01A1058372-05) and Wellcome Trust-Burroughs tion in the cohort study and, of this group, about half of them Wellcome Fund (059113/Z/99/A & 059113/Z/99/Z) grants. The achieved sustained viral response (SVR) [27]. The relatively short cohort in Aghour El Soughra was funded by the Hepatitis C mean length of follow-and the two possible misclassification Prevention Project, USAID grant 263-G-00-96-00043-00. errors may mean that the effect of chronic HCV infection on mortality may be slightly underestimated. In addition, the effect of some HCV transmission risk factors Author’s contributions that have been included in previous studies were not assessed in our study: injecting drug use, alcohol abuse, and co-infection Mostafa K Mohamed, Arnaud Fontanet and Mohamed Abdel- with HIV or HBV [13,14,16]. However, these are unlikely to con- Hamid initiated the cohorts and Arnaud Fontanet, Mostafa K found the association between HCV infection and mortality in Mohamed and Gamal Esmat were responsible for the design of rural Egypt, where injecting drug use, alcohol abuse, and HIV the current study; Ahmed Medhat, Iman Bakr, and Naglaa Arafa infection are very rare or even non-existent [34,35]. Although for fieldwork; Mai El-Daly and Mohamed Abdel-Hamid for labo- HBV is endemic in Egypt (prevalence of chronic HBV infection: ratory assays; and Mostafa El Hoseiny, Nabiel Mikhail, Chesnais 1.7% in general population [36]), and shares the same route of Cédric, Aya Mostafa, Yusuke Shimakawa and Arnuad Fontanet transmission with HCV, its prevalence is similar between people for data and statistical analysis. Aya Mostafa and Yusuke Shi- with and without HCV infection in rural Egypt [34]. makawa drafted the manuscript, and all the authors reviewed Overall, the all-cause mortality rate attributable to chronic and approved it. HCV infection was 5.7% (95% CI: 1.0–10.1%), while liver-related mortality attributable to chronic HCV infection was 45.5% Acknowledgements (11.3–66.4%). This has important implications for public health policy and action. In 2006, the Egyptian Ministry of Health and We would like to thank Adeline Bernier, Mohand Aït-Ahmed, Population established the National HCV Treatment Programme. Yoann Madec, and Rebecca Grant for their technical advice, the Since then, more than 200,000 people have been treated with people of the three villages in rural Egypt for their participation 48 weeks of PegIFN and ribavirin [5]. In 2014, following the in this study and the field teams responsible for data collection. agreement between the Egyptian Government and Gilead for the purchase of sofosbuvir at the generic price of $900 (U.S.) for References 12 weeks, a new treatment regimen including sofosbuvir was started, and is expected to achieve SVR in >90% of patients [37]. [1] Gower E, Estes CC, Hindman S, Razavi-Shearer K, Razavi H. Global epidemi- Assuming that SVR rate is 90% with the new regimen and patients ology and genotype distribution of the hepatitis C virus infection. J Hepatol who achieved SVR have a similar mortality to people without 2014;61:S45–S57. http://dx.doi.org/10.1016/j.jhep.2014.07.027. chronic HCV (i.e., never infected or spontaneously cleared HCV), [2] GBD 2013 Mortality and Causes of Death Collaborators. Global, regional and national levels of age-specific mortality and 240 causes of death, 1990– treating all the villagers with positive HCV RNA may reduce all- 2013: a systematic analysis for the Global Burden of Disease Study 2013. cause mortality rate by up to 5% and hepatic mortality rate by Lancet 2015;385:1990–2013. http://dx.doi.org/10.1016/S0140-6736(14) up to 40% in rural Egypt, on the basis of the PAFs obtained in 61682-2. our study. Still, due to the huge reservoir of HCV-infected individ- [3] Guerra J, Garenne M, Mohamed MK, Fontanet A. HCV burden of infection in uals in Egypt, it will take many years until the impact of the Egypt: results from a nationwide survey. J Viral Hepat 2012;19:560–567. http://dx.doi.org/10.1111/j.1365-2893.2011.01576.x. National Treatment Programme and other control initiatives on [4] Pawlotsky JM. New hepatitis C therapies: the toolbox, strategies, and HCV-related mortality can be documented. challenges. Gastroenterology 2014;146:1176–1192. http://dx.doi.org/ In summary, this study described the excess mortality rates 10.1053/j.gastro.2014.03.003. associated with HCV infection among young and middle-aged [5] Center for Disease Control and Prevention (CDC). Progress toward preven- men in high prevalence communities of rural Egypt. This is the tion and control of hepatitis C virus infection–Egypt, 2001–2012. MMWR Morb Mortal Wkly Rep 2012;61:545–549. first study of this kind to be conducted in a resource-limited set- [6] Osella AR, Misciagna G, Guerra VM, Chiloiro M, Cuppone R, Cavallini A, et al. ting. These results should be used to inform the prioritisation of Hepatitis C virus (HCV) infection and liver-related mortality: a population- patients in the Egypt National Treatment Programme which uses based cohort study in southern Italy. Int J Epidemiol 2000;29:922–927. a new highly potent antiviral agent. [7] Amin J, Law MG, Bartlett M, Kaldor JM, Dore GJ. Causes of death after diagnosis of hepatitis B or hepatitis C infection: a large community-based linkage study. Lancet 2006;368:938–945. http://dx.doi.org/10.1016/S0140- 6736(06)69374-4. Conflict of interest [8] Neal KR, Ramsay S, Thomson BJ, Irving WL. Excess mortality rates in a cohort of patients infected with the hepatitis C virus: a prospective study. Gut 2007;56:1098–1104. http://dx.doi.org/10.1136/gut.2006.113217. The authors who have taken part in this study declared that they [9] Duberg AS, Törner A, Davisdóttir L, Aleman S, Blaxhult A, Svensson Å, et al. do not have anything to disclose regarding funding or conflict of Cause of death in individuals with chronic HBV and/or HCV infection, a interest with respect to this manuscript. nationwide community-based register study. J Viral Hepat 2008;15:538–550. http://dx.doi.org/10.1111/j.1365-2893.2008.00982.x.

Journal of Hepatology 2016 vol. 64 j 1240–1246 1245 Research Article

[10] Prasad L, Spicher VM, Negro F, Rickenbach M, Zwahlen M, Negro F, et al. population description and HCV prevalence. Hepatology 2000;32:111–115. Little evidence that hepatitis C virus leads to a higher risk of mortality in the http://dx.doi.org/10.1053/jhep.2000.8438. absence of cirrhosis and excess alcohol intake: the Swiss Hepatitis C Cohort [24] Habib M, Mohamed MK, Abdel-Aziz F, Magder LS, Abdel-Hamid M, Gamil F, Study. J Viral Hepat 2009;16:644–649. http://dx.doi.org/10.1111/j.1365- et al. Hepatitis C virus infection in a community in the Nile Delta: risk factors 2893.2009.01113.x. for seropositivity. Hepatology 2001;33:248–253. http://dx.doi.org/10.1053/ [11] Uto H, Stuver SO, Hayashi K, Kumagai K, Sasaki F, Kanmura S, et al. Increased jhep.2001.20797. rate of death related to presence of viremia among hepatitis C virus [25] Mostafa A, Taylor SM, El-Daly M, El-Hoseiny M, Bakr I, Arafa N, et al. Is the antibody-positive subjects in a community-based cohort study. Hepatology hepatitis C virus epidemic over in Egypt? Incidence and risk factors of new 2009;50:393–399. http://dx.doi.org/10.1002/hep.23002. hepatitis C virus infections. Liver Int 2010;30:560–566. http://dx.doi.org/ [12] McMahon BJ, Bruden D, Bruce MG, Livingston S, Christensen C, Homan C, 10.1111/j.1478-3231.2009.02204.x. et al. Adverse outcomes in Alaska natives who recovered from or have [26] Breban R, Arafa N, Leroy S, Mostafa A, Bakr I, Tondeur L, et al. Effect of chronic hepatitis C infection. Gastroenterology 2010;138:922–931. http:// preventive and curative interventions on hepatitis C virus transmission in dx.doi.org/10.1053/j.gastro.2009.10.056. Egypt (ANRS 1211): a modelling study. Lancet Glob Heal 2014;2:e541–e549. [13] Omland LH, Krarup H, Jepsen P, Georgsen J, Harritshøj LH, Riisom K, et al. http://dx.doi.org/10.1016/S2214-109X(14)70188-3. Mortality in patients with chronic and cleared hepatitis C viral infection: a [27] El Makhzangy H, Esmat G, Said M, Elraziky M, Shouman S, Refai R, et al. nationwide cohort study. J Hepatol 2010;53:36–42. http://dx.doi.org/ Response to pegylated interferon alfa-2a and ribavirin in chronic hepatitis C 10.1016/j.jhep.2010.01.033. genotype 4. J Med Virol 2009;81:1576–1583. http://dx.doi.org/10.1002/ [14] Omland LH, Christensen PB, Krarup H, Jepsen P, Weis N, Sørensen HT, et al. jmv.21570. Mortality among patients with cleared hepatitis C virus infection compared [28] Abdel-Hamid M, Edelman DC, Highsmith WE, Constantine NT. Optimization, to the general population: a danish nationwide cohort study. PLoS One assessment, and proposed use of a direct nested reverse transcription- 2011;6. http://dx.doi.org/10.1371/journal.pone.0022476. polymerase chain reaction protocol for the detection of hepatitis C virus. J [15] Grebely J, Raffa JD, Lai C, Kerr T, Fischer B, Krajden M, et al. Impact of Hum Virol 1997;1:58–65. hepatitis C virus infection on all-cause and liver-related mortality in a large [29] Newson RB. Attributable and unattributable risks and fractions and other community-based cohort of inner city residents. J Viral Hepat scenario comparisons. Stata J 2013;13:672–698. 2011;18:32–41. http://dx.doi.org/10.1111/j.1365-2893.2010.01279.x. [30] Freeman AJ, Dore GJ, Law MG, Thorpe M, Von Overbeck J, Lloyd AR, et al. [16] Lee MH, Yang HI, Lu SN, Jen CL, You SL, Wang LY, et al. Chronic hepatitis C Estimating progression to cirrhosis in chronic hepatitis C virus infection. virus infection increases mortality from hepatic and extrahepatic diseases: a Hepatology 2001;34:809–816. http://dx.doi.org/10.1053/jhep.2001.27831. community-based long-term prospective study. J Infect Dis [31] D’Amico G, Garcia-Tsao G, Pagliaro L. Natural history and prognostic 2012;206:469–477. http://dx.doi.org/10.1093/infdis/jis385. indicators of survival in cirrhosis: a systematic review of 118 studies. J [17] Thursz M, Fontanet A. HCV transmission in industrialized countries and Hepatol 2006;44:217–231. http://dx.doi.org/10.1016/j.jhep.2005.10.013. resource-constrained areas. Nat Rev Gastroenterol Hepatol 2014;11:28–35. [32] Davis GL, Alter MJ, El-Serag H, Poynard T, Jennings LW. Aging of hepatitis C http://dx.doi.org/10.1038/nrgastro.2013.179. virus (HCV)-infected persons in the United States: a multiple cohort model [18] Frank C, Mohamed MK, Strickland GT, Lavanchy D, Arthur RR, Magder LS, of HCV prevalence and disease progression. Gastroenterology 2010;138. et al. The role of parenteral antischistosomal therapy in the spread of http://dx.doi.org/10.1053/j.gastro.2009.09.067 e6. hepatitis C virus in Egypt. Lancet 2000;355:887–891. http://dx.doi.org/ [33] Hajarizadeh B, Grebely J, Dore GJ. Epidemiology and natural history of HCV 10.1016/S0140-6736(99)06527-7. infection. Nat Rev Gastroenterol Hepatol 2013;10:553–562. http://dx.doi. [19] Arafa N, El Hoseiny M, Rekacewicz C, Bakr I, El-Kafrawy S, El Daly M, et al. org/10.1038/nrgastro.2013.107. Changing pattern of hepatitis C virus spread in rural areas of Egypt. J Hepatol [34] Darwish MA, Faris R, Darwish N, Shouman A, Gadallah M, El-Sharkawy MS, 2005;43:418–424. http://dx.doi.org/10.1016/j.jhep.2005.03.021. et al. Hepatitis c and cirrhotic liver disease in the Nile delta of Egypt: a [20] Nafeh MA, Medhat A, Shehata M, Mikhail NN, Swifee Y, Abdel-Hamid M, community-based study. Am J Trop Med Hyg 2001;64:147–153. et al. Hepatitis C in a community in Upper Egypt: I. Cross-sectional survey. [35] National AIDS Program. Global AIDS Response and Progress Report 2014, Am J Trop Med Hyg 2000;63:236–241. Egypt. 2014. [21] Medhat A, Shehata M, Magder LS, Mikhail N, Abdel-Baki L, Nafeh M, et al. [36] Schweitzer A, Horn J, Mikolajczyk RT, Krause G, Ott JJ. Estimations of Hepatitis c in a community in Upper Egypt: risk factors for infection. Am J worldwide prevalence of chronic hepatitis B virus infection: a systematic Trop Med Hyg 2002;66:633–638. review of data published between 1965 and 2013. Lancet 2015;6736:1–10. [22] Mohamed MK, Abdel-Hamid M, Mikhail NN, Abdel-Aziz F, Medhat A, Magder http://dx.doi.org/10.1016/S0140-6736(15)61412-X. LS, et al. Intrafamilial transmission of hepatitis C in Egypt. Hepatology [37] Asselah T. Optimism for patients with genotype 4 HCV infection: clinical 2005;42:683–687. http://dx.doi.org/10.1002/hep.20811. trials with direct-acting antivirals ﬁnally available. J Hepatol [23] Abdel-Aziz F, Habib M, Mohamed MK, Abdel-Hamid M, Gamil F, Madkour S, 2015;62:996–999. http://dx.doi.org/10.1016/j.jhep.2015.03.003. et al. Hepatitis C virus (HCV) infection in a community in the Nile Delta:

1246 Journal of Hepatology 2016 vol. 64 j 1240–1246